One advantage these chemical systems may have over compressed tanks is how well they are suited for intermittent use and long-term storage (as opposed to regular, continuous operation). The US FAA Aviation Maintenance Technician Handbook points out this advantage on page 16-5:
Sodium chlorate chemical oxygen generators also have a long shelf
life, making them perfect as a standby form of oxygen. They are inert
below 400 °F and can remain stored with little maintenance or
inspection until needed, or until their expiration date is reached.
Compressed gas cylinders, on the other hand, require regular maintenance and testing whether in use or in storage. Pressure vessels are prone to metal fatigue and corrosion over extended periods. Exposure to extreme temperatures can accelerate fatigue or cause an already-weakened cylinder to fail; COG systems, in comparison, are somewhat less vulnerable to temperature extremes.
As part of an emergency system, an oxygen cylinder would experience infrequent use and mechanical fatigue would be less of a concern than corrosion.* Corrosion is driven mainly by the reactivity and partial pressure of the stored gas and by the presence of moisture in the cylinder. Proper selection of fittings such as valves and regulators is critical to minimize the potential for a galvanic reaction between the fitting and the cylinder.
Keep in mind that pressure vessels support a difference of pressure between the inside and the outside of the vessel. During a decompression event, the outside pressure drops rapidly while the inside pressure stays constant, leading to a rapid increase in stress on the vessel. This has two implications: first, the capacity of the tank must be reduced to account for the lower atmospheric pressure at operating altitude; second, a worn tank is most likely to explode just when you need it the most—not at all a desirable quality in a safety device. (COG systems do not, to my knowledge, explode, though they can start fires when handled improperly.)
This leads to strict inspection and maintenance requirements for compressed gas cylinders. They have to be rotated in and out of service every so often to go through requalification testing and potentially reconditioning or condemnation. For some idea of how complicated these requirements are, take a look at this 2003 Aircraft Maintenance Technology article.
All of this, of course, represents a cost to the operator—service contracts, training, etc. In an industry that competes mainly on ticket prices, cost is king, and reduced maintenance costs would be an attractive selling point.
It's also nice to be able to relate the safety system cost directly to the occasions when you need to deploy the system. Going back to the initial point about suitability for intermittent, emergency applications: You're paying maintenance costs for that cylinder whether there's a depressurization or not. The cost of the oxygen itself is probably minor in comparison. So reducing the frequency of incidents by an order of magnitude may not change the cost very much. On the other hand, only paying for the COG system when you use it means the airline can reduce its costs by reducing depressurization incidents. This isn't so much an engineering concern, and I am speculating a bit here; the point is, businesses often don't make decisions in the same way that engineers do, and that's something to keep in mind when comparing design alternatives.
* You might think that pure oxygen is not a corrosive gas (perhaps because we breathe oxygen) but the definition of "corrosive" can vary a bit with context. For example, the Air Liquide Design and Safety Handbook (p. 2) uses this definition:
These are gases that corrode material or tissue on contact, or in the
presence of water. ... Due to the probability of irritation and damage
to the lungs, mucous membranes and eye tissues from contact, the
threshold limit values of the gas should be rigidly observed. Proper
protective clothing and equipment must be used to minimize exposure to
The emphasis here is on personal safety and health impacts and in that context, oxygen isn't a corrosive gas. But in the context of storing a gas in a pressurized cylinder, the emphasis is on the potential for the stored gas to react chemically with the walls or components in a way that weakens them, and in that context, pure oxygen definitely counts as "corrosive."